MCP

Pythonic framework for building MCP servers. Decorators for tools, resources, and prompts. Officially incorporated into the Python SDK.

Bridge that exposes stdio MCP servers over HTTP/SSE, or vice versa. Run a local server remotely, or connect to a remote server as if it were local.

Configure model, effort, tools, MCP servers, and invocation scope.

Configure model, tools, MCP, skills, memory, and scoping.

Route specific MCP servers only to specific subagents.

Fires when an MCP server requests input from the user.

Connect external tools and data sources via the open MCP standard.

Local, project, user, and plugin-level MCP configurations.

Deferred tool loading reduces context overhead for large MCP suites.

Pre-configured or dynamic OAuth for remote MCP servers.

Reference and read resources exposed by MCP servers.

Execute MCP prompts as commands via the slash menu.

Receive push messages from MCP servers via channels.

Admin-controlled allow and deny lists for MCP servers.

Most MCP servers are noise. After shipping 24 apps with Claude Code, these are the five I reach for every time.

MCP is the USB-C of AI agents. What the Model Context Protocol is, why Anthropic built it, and how to install your first server in Claude Code or Cursor. Fact-checked against the official MCP spec.

MCP isn't just a plugin format - it's a full JSON-RPC protocol for connecting LLMs to tools, resources, and prompts. Here's how it works under the hood, sourced from the official spec.

Hacker News keeps arguing about Claude Code, Codex, skills, MCP, and orchestration. Under the noise, the same four truths keep surfacing: workflows matter more than demos, verification is the bottleneck, skills beat prompts, and orchestration matters more than raw autonomy.

The coding-agent workflow is maturing past giant hand-written prompts. The winning pattern in 2026 is a control stack: project rules, reusable skills, bounded sub-agents, and deterministic tools around the model.

An opinionated guide to the MCP server ecosystem in 2026. Curated picks by category, real configuration examples, installation commands, and honest assessments of what works and what does not.